In bioinformatics, the database of biological sequences increases at a dizzying rate, with the alignment algorithms used for the comparison of sequences determining genetic distances, generation of phylogenetic trees, etc. This work seeks to compare the incorporation of the Rabin-Karp and Base 5 algorithms as possible optimizations during the generation of seed indexes of the BLAST alignment algorithm to align multiple query sequences with the DNA sequence of the human genome as sequence of reference. The tests were processed sequentially and using GPU in the MANATI supercomputer of the High Performance Computational Center of the Peruvian Amazon of the IIAP, showing a better performance for a possible optimization of BLAST in the generation of hash keys with the algorithm taken from Base 5 for long sequences (genomes) with short keys, generating maximum dispersion. However, for short sequences or longer keys, it is advisable to use Karp-Rabin, reducing this dispersion.